I get the story. I even kind of like it, but I still can't help thinking it would have been cheaper just to outfit everything with toilets and get on with life.

I kind of feel the same way about laptop "stand by" for energy saving (good motive, wasteful execution). My time = money and the cost of having my laptop turned on needlessly per hour is less than 2 cents while my wasted time every time it happens when I'm just about to use it/reading what it says is worth more than 2 cents. If it is about the environment I would rather them just double the price of electricity and use that towards green initiatives.

The idea is that, initially, all cars had toilets (see the invisible emphasis of "only" in "if only fifty percent of the cars would be equipped with a toilet"). Of course, there must have been plenty of cost associated with handling toilet complaints and training (and re-training) at the shunting yard, but we can guess that the long term savings made up for that.

I think the joke is that after sticking two cars together for all eternity (one with a toilet, one without), they have effectively become one car. So the original solution has been restored: every car has a toilet.

just for the record: there are lots of brilliant articles in dijkstra's EWDs. i am particularly fond of trip reports (almost always funny), programming language comments (e.g. EWD498, or EWD898: "Stop BASIC before it stops you") and insights to the commercial field of computer science (EWD 898: "I have fond memories of a project of the early 70's that postulated that we did not need programs at all! All we needed was intelligence amplification. If they have been able to design something that could amplify at all, they have probably discovered it would amplify stupidity as well; in any case I have not heard from it since.")

Might he have been talking about Douglas Englebart's "Augmenting Human Intellect" project? (which invented the mouse as well as provided the foundation of GUIs, hypertext, and the Internet -- although any amplification of intelligence is debatable.)

The children's bedtime story [from a former(?) Russian satellite] version of this parable is a projector film slide series titled "The coloured pencil", from the same era. It cheekily teaches the parents on how a "cost driven market based" pencil factory had a brilliant innovator, arguing, that there is no need for graphite in the end of a pencil, since nobody can use pencils when they are shorter than a cm or so. The savings made on using less graphite was stellar, the innovator got promoted.

Seeing the success, soon the second innovation followed: what use is the wood at the end of the pencil, when there isn't any graphite inside? Less wood, more profit.

And this handling of "externalised" costs continued until the pencil became dysfunctional, unusable (Dijkstra as a programmer would notice the powerful idea of recursion).

The story ends with the reinvention of the pencil, and a huge prime for the genius.

Back to Dijkstra's version: an MBA market analyst would come up with a "research" study which concludes, that in order to improve the "user experience" of the customers on the train, at least (two) WC-s would be needed in each wagon, pocketing a hefty consulting fee for the ground breaking innovation.